Application of 2007 NESC Arc-Flash Requirements 2

[dpc]

I'm curious how others are dealing with the 2007 NESC arc-flash protection requirements and the apparent contradictions with existing NFPA 70E?

In particular, I'm wondering about the provisions for dealing with systems less than 1000 V. The NESC says that in lieu of doing an actual arc-flash study, a hazard of 4 cal/cm2 can be assumed under 1000 V. For 480 V systems, this can be absurdly low when compared with the requirements of NFPA 70E and IEEE 1584. This is especially true when dealing with the 480V portion of a pad-mounted transformer that probably has minimal primary protection.

Does anyone know the basis or origin of the 4 cal/cm2??

Also, for substations using metal-clad switchgear, the IEEE 1584 calculations would seem to be much more relevant than the NESC tables when dealing with switchgear. Anyone made any decisions on this?

 

[stevenal]

dpc,

Have you seen the change proposals? Here is what the working group said:

"The working group rejected the original CPs and agreed to develop a revised CP that would address worker
safety associated with electric arc exposure, while at the same time acknowledging industry needs and unique
working conditions. It was recognized by the Working Group that rule changes in Section 41 for employers,
and Section 42 for employees, were necessary to accomplish this task. This CP focuses on those issues and
will provide work rules that are appropriate for performing work associated with electric supply and communication
systems."

I think we are seeing the compromise between those who are looking at the potential hazard and those who are looking at the actual safety record of the industry. Some of the dissenting comments are interesting too.

I agree that the arc in a box method of IEEE is more relevant for metal clad than the open air single phase fault that ArcPro (and NESC tables) use.

 

[dpc]

Steven,

Thanks - are the Change Proposals accessible on-line? I am interested in figuring out how this came about. I suspect you're right - as usual a compromise.

Dave

 

[stevenal]

No online version that I am aware of. Should be able to purchase and download the pdf from IEEE.

 

[apowerengr]

dpc,

So far I've structured my approach using the NESC tables for live-line work, ArcPro for work in the area but not on energized equipment (sub inspections and the like) and use PTW w/ either NFPA 70E or IEEE 1584 methods for the < 1000 volt work. You can't use the tables < 1000 volts and the 4 cal/cm2 alternate is absurd if you look at calculated results. Switchgear configuration for padmounts and panel for metering sockets? Trying to catch work performed by meter technicians, locators, tree-trimmers, etc. with the appropriate working exposures.

I agree with you on using IEEE 1584 for metal-clad gear and MV faults although I don't believe the commercially available software is applicable for MV systems? I've also started conversations with Cooper Power Systems on their standard bayonet fusing charts and the resulting exposures for padmounted applications looking for a different fuse or different sizing.

When using the NESC table for live-line work are you just considering the total clearing time of the primary protective device or the backup device per IEEE 1584? How do you handle intermediate fault values, especially between 0 and 5kA?

I have some utility customers encountering conflicts with installations where utility equipment is located in customer switchgear - which code applies? The Utility says today they only need FR or to prove cotton won't combust - the facitlity owner says the utility must meet the Owner's rated FR standard under NFPA 70E. The utility does not have FR rated to the 70E calculated values so we countered with requiring a plant outage - not well received.

I think this new code needs a lot of discussion - have you considered requesting any interpretations from the code committee?

Dave

 

[dpc]

apowerengr,

Thanks for the input. I think we're both seeing this in basically the same way.

Just a couple of points:

I thought the tables in NESC were computed using ArcPro, so for similar conditions and assumptions, ArcPro calcs should match the NESC table?

IEEE-1584 has methods for calculation medium-voltage arc-flash. Above 15 kV, I believe they simply use the Lee equations, but below 15 kV, there are specific equations in 1584. We use EasyPower and it calculates medium-voltage (and above) arc-flash levels using IEEE-1584. This still seems appropriate to me for metal-clad switchgear, even in utility substation.

We don't consider backup protection (breaker failure) per se in doing arc-flash calculations per NFPA 70E and IEEE 1584. For main breakers you do have to use the next device upstream to determine fault duration.

I can (sort of) see some logic to the 4 cal/cm2 value for 120/240V and 208V systems, but at least for 480V and above, we will use the IEEE 1584 calculations.

 

[wbd]

dpc,

"For main breakers you do have to use the next device upstream to determine fault duration."

But that would be only if you need the AFH on the line side of the breaker, say between the low side of the transformer to the line side of the main breaker, Correct?

Also are you using the CLF feature in EasyPower?

 

[dpc]

wbd,

Right - for any work on the main breaker compartment or section, you would have to assume the fault could be on the line side of the breaker, so the fault would have to be cleared by the next upstream device.

I generally don't bother with the CLF option in EasyPower. Just based on time of clearing, the arc-flash becomes so low that the difference isn't worth worrying about. The big problem with fuses is that often the fault current is not high enough to get into the current-limiting range of the fuse. This is especially true when the fuse is on the primary of a transformer and fault is on the secondary side.

 

[Breakerboy]

We recently performed new calcs that include each component in the system using the exact equipment we have in our system. Our calcs show that even our 489v load centers are in the 20cal/cm squared levels. The NFPA 70E and other arc flash software are not using real values for specific equipment. We are installing remote breaker racking on all of our 13.8kv and 4.16kv breakers and when that is finished we are going to install the systems on our 480v load centers. We decided that once we knew the real hazards we wouldn't take chances with our employees. Check out the Feb issue of Power Magazine for a very good problem defining article which Dr. Peltier put together with IEEE and NFPA input.

 

[Eleceng01]

Breakerboy, do you have a link to that article? I can't find it. Thanks in advance.

 

[Breakerboy]

http://www.powermag.com/ExportedSite/Archives/Archives.htm

They did an accompanying article on our project that is linked in the first paragraph or so under (see box).

 

[JensenDrive]

Is the present art of arc flash analysis giving enough credit for those that have safe practices and new well maintained robust equipment? The present art says, "If something goes wrong, here is the arc level." Sure, a valid concern, but what about some analysis of the probability of something going wrong? For instance, I see several comments about the arc flash case where the bus main breaker taken out of the picture, and the xfmr high side fuse needs to clear the fault. How likely is the main breaker to fail, or how likely is someone to work on wires upstream of the main breaker, live? Where does this fit into the standards?

 

[Eleceng01]

JesenDrive, wouldnt opening/closing the main breaker be considered interacting with the device, hence, the need for proper PPE. The main breaker could fail during the opening or closing and cause a arc flash. Imagine the case of racking in the main breaker with it accidentally closed. I read an artcle recently of this type of probelm where unfortunately 2 were killed.

 

[dpc]

It's true that some failures are less likely than others. But a main breaker is just as likely to fail as any other breaker.

If your logic is that we don't need to worry about PPE because we have a very reliable system and we know what we are doing, you aren't going to get very far in discussions with OSHA, etc.

You can separate it into two categories: Hazard and Risk.

For arc-flash, the **hazard** is determined by the worst case arc energy that can be released at a particular point in the system, based on the physical location, the fault current and the clearing time. The hazard does not change unless you change the system. The **risk** is a function of what is being done. Someone just opening a cubicle door is at a much lower risk than someone who is racking a breaker into an energized bus.

But if there is a fault, the energy released would be basically the same. So does it make sense to allow someone just opening a door to wear a lower level of PPE than someone racking in a breaker for the same HAZARD level. That's the question. It's a valid question and a legitimate position. I don't agree with it, but others support it.

NFPA 70E 2004 does not provide a mechanism for this, but I'm sure it under consideration in the upcoming revision. Stay tuned.

 

[Breakerboy]

I would like to comment on the question about probability and risk. One of the things we have to remember is that there are 5-10 arc flash incidents in the US each day. We also have to remember that even the newest and best components can fail. We were just months away from our plant start up and were running 13.8kv cooling pumps and had a fault develop at our Circulating Water Screen house. The fault propagated back to the main bus in the Turbine building. The fault blew the 200lb door off of the cubicle, ripping the 3- 1/8"hinges, the doorstop and the two 3/16" latches off with it. The door traveled 25' until it crashed into the adjacent bus. 30' away at a right angle to the cubicle were three test breakers still in their wooden shipping crates. The Blast charred the boxes. Eyewitnesses saw the fireball travel the 25'across with the door and then up 16' to 20' where it burned through rope that was holding up a weld cable. Copper shrapnel was sprayed for 40-50' in an arc that started at 40- degree angle out of the cubicle. Our operator had just walked away from the area 10 seconds before the blast. New breakers, New pumps, New relays, New everything and we could have been attending his funeral. This is one thing that won't result in a cut or a minor burn or a broken bone, this is something that will result in death or permanently maimed or burned employee. Don't be lulled into believing that your equipment is somehow immune to failure.

 

[rbulsara]

breakerboy:

Just to add to your point:

The current arc flash analysis standards and recommendations for PPE are ONLY to minimize "burn" induries due to arc-flash. All other common sense precautions to safe guard against flying debris and shrapnels are still to be implemented and are not and nor can they be covered by code books.

Any class of PPE in cases like the one described by you are of no help.

 

[Breakerboy]

That is why when we learned of the blast effects and determined our cal/cm2 energy levels, we installed remote breaker racking. We looked into everything on the market and found that most of the systems were simply remote switches that had slight differences in how they did the job. We realized that if you are not watching the breaker during the racking process, mechanical malfunctions could result and without the operator there to stop the process, damage could occur that could cause arc flash blasts to occur or at best additional damage to the equipment. We finally chose Switchgear Solutions system called Safe-T-Rack. It provides equipment protections that give you some assurances in case of binding, chain breaks, alignment issues etc. Their system has a component called "LimiTilt" that stops the racking motor if the breaker becomes out of level either fore and aft or side to side. It also has an overcurrent device that will stop the racking motor if an overcurrent situation occurs due to binding or if the shutters do not open etc. This allows the operator to be at least 30' away from the breaker and in most cases up to 50', and at a right angle to the blast opening. I also allows the door to be latched which minimizes the boundary area. The door absorbs some of the blast energy as it deforms. In our investigations all of the major breaker arc flash blast come out the front of the cubicle with some bowing of the side walls. By moving the operator away from the blast angle and out of the heat area we felt that we were providing the best PPE and equipment protection on the market. We are installing this on 90 13.8kv and 4.16kv breakers. Installation takes less than 2 hours per cubical and once your team is experienced on a few they can cut that down to 1.5 hours. The company will come to your plant and demonstrate it for you and give you opportunity to test it out. They are can do people with the ability to think outside of the box.

 

[Grollor33]

Hi all,

I see like a lot of good discussion, but I think I missed the answer to the original problem...

Is it a good approx to assume 4 cal/cm^2 for <1000V systems??

The calculations for the arc energy are very dependent on the arc duration...

Has anyone seen test data where they've actually been able to get a LV arc to sustain long enough to get the calculated energy exposure?

 

[dpc]

It's important to know exactly what voltage is under discussion. The term "low voltage" is much too broad when discussing arc-flash energy.

4 cal/cm2 is much too low for a 480 V system. It might be a reasonable starting point for a residential 120/240 V system, but I don't know of anyone who think it makes sense for 480 V. I think the NESC committee just punted on this one.

480 V and 277 V arcs can be sustained for a long, long time. At 240 V it is much harder to sustain an arc, and even more difficult (but not impossible) at 208 V and 120 V.

IEEE 1584 has a method for calculating approximate arc-flash energies at voltages to 15 kV based on actual test data. 480 V has tremendous arc-flash potential, but arcing currents will be significantly reduced from bolted fault currents.

 

[rbulsara]

Jweave33:

The current arc flash standards and guidlines (IEEE 1584 and NFPA 70E) are only intened to safeguard person workig live on it, so the maximum time to be considered is 2 seconds. The thinking is that by that time either the person is removed from the fault or is dead or is entangled beyond protection. So it does not matter how long the arc persists. The arc flash energy calculated in the total accumulated for the entire event but only long enough as it pertains to safety of the personnel NOT the equipment.